This project began with the acquisition of short-wavelength, high-dispersion IUE spectra of selected late O- and early B-type stars that are near the main sequence in open clusters and associations. The profiles of the resonance lines of N(V), Si(IV), and C(IV) were studied, and we found that the C(IV) lines are the most sensitive indicators of mass loss (stellar winds) in stars of this type. The mass loss manifests itself as an extension of the short-wavelength absorption wing of the doublet, while there is no P Cygni-type emission on the long-wavelength side of the line profile. We investigated whether the short-wavelength extension could be caused by blended lines of other ionic species formed in the photosphere. Although blending is present and introduces uncertainty into the estimation of the precise location on the main sequence of the onset of the mass-loss signature, it is a crucial issue only in a few marginal cases. Mass loss certainly overwhelms blending in its influence on the spectrum between spectral types B0 and B1 (effective temperatures in the range 25,000-27,000 K). We defined a parameter called P(sub w), to describe the degree of asymmetry of the C(IV) resonance-line profile, and we studied the dependence of this parameter on the fundamental stellar parameters. For this purpose, we derived new estimates of the stellar T(eff) and log g from a non-LTE, line-blanketed model-atmosphere analysis of these stars (Grigsby, Morrison, and Anderson 1992). In order to estimate the stellar luminosities, we performed an exhaustive search of the literature for the most reliable available estimates of the distances of the clusters and associations to which the program stars belong. The dependence of P(sub w) on stellar temperature and luminosity is also studied.